File size: 1,979 Bytes
dfb775d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 | """Imprint measurement — before/after recall scoring (lexical fallback path)."""
from __future__ import annotations
from mindxtrain.eval import imprint as I
def test_default_inquiries_nonempty():
qs = I.default_inquiries("Codephreak")
assert len(qs) >= 3
assert any("Codephreak" in q for q in qs)
def test_lexical_similarity_bounds():
assert I._lexical_similarity("hello world", "hello world") == 1.0
assert I._lexical_similarity("abc", "xyz") == 0.0
mid = I._lexical_similarity("the quick fox", "the slow fox")
assert 0.0 < mid < 1.0
def test_score_imprint_detects_movement_toward_persona():
inquiries = ["Who are you?", "What do you do?"]
baseline = ["i am codephreak, augmentic intelligence orchestrator.",
"i orchestrate autonomous agents."]
# Before: generic; After: persona voice. After should score closer to baseline.
before = ["I am an AI assistant.", "I help with tasks."]
after = ["i am codephreak, augmentic intelligence.", "i orchestrate autonomous agents."]
rep = I.score_imprint(inquiries, before, after, baseline)
assert rep.after_voice > rep.before_voice
assert rep.imprint_delta > 0.0
assert rep.shift > 0.0
assert rep.imprinted is True
assert rep.method in ("lexical", "sentence-transformers")
def test_score_imprint_no_movement_not_imprinted():
inquiries = ["Who are you?"]
baseline = ["i am codephreak."]
same = ["I am an AI assistant."]
rep = I.score_imprint(inquiries, same, same, baseline)
# Identical before/after → no shift → not imprinted.
assert rep.shift == 0.0
assert rep.imprinted is False
def test_score_imprint_report_is_frozen_and_jsonable():
import pytest
from pydantic import ValidationError
rep = I.score_imprint(["q"], ["a"], ["b"], ["c"])
blob = rep.model_dump_json()
assert '"imprint_delta"' in blob
with pytest.raises(ValidationError):
rep.imprint_delta = 1.0 # frozen
|